Physician, Attorney, Journalist
San Antonio, TX
Joined March 2015
- Tweets10.5K
- Following1.6K
- Followers2.1K
- Likes12.6K
Brant Mittler retweeted
I’m at the airport. Of course I start reading weekend edition of the @WSJ.
Wouldn’t you know it, the dual designated non-profit grifter, NYU Langone, has a full page ad.
And you’re still wondering why healthcare is expensive?
And you know that means taxpayers are buying WSJ ads for NYU Langone.
Taxpayers subsidize its land.
Taxpayers subsidize its buildings.
Taxpayers subsidize its debt.
Taxpayers subsidize its expansion.
What did it cost? The Wall Street Journal rate card prices a national full-page color advertisement at more than $300,000 (52 weeks).
For one page.
This is the same “nonprofit” hospital system that advertised on a Super Bowl commercial.
Its hospital enterprise reports nearly $10 billion in property, plant, and equipment. It has more than 6,000 physicians.
It operates more than 330 outpatient locations. It buys hospitals. It buys orthopedic groups. It buys physician practices.
It buys land.
It buys advertising.
It buys the commercially insured patients every independent physician needs to survive.
Then it calls itself a charity.
How many taxpaying neurosurgery practices were destroyed while NYU Langone built this empire?
How many physicians were forced to sell?
How many became employees of the institution they once competed against?
How many independent practices had their referrals siphoned away, their patients captured, and their businesses slowly suffocated by a competitor taxpayers were forced to subsidize?
Nobody knows. Nobody counts the bodies. An independent practice dies quietly.
First, referrals disappear.
Then revenue collapses.
Then employees are fired.
Then the physician sells.
Then the hospital raises the price, adds a facility fee, and hangs its logo over the corpse.
The independent physician paid income taxes, property taxes, payroll taxes, commercial rent, malpractice premiums, and the full cost of building a practice.
NYU Langone took the tax exemption, accumulated billions in assets, swallowed the competition, and bought an advertisement congratulating itself.
Where is the charity?
NYU Langone wants the power of a corporation, the market share of a monopoly, the real estate portfolio of a conglomerate, and the tax treatment of a church.
No.
If you acquire competitors like a corporation, advertise like a corporation, accumulate billions like a corporation, and hunt profitable patients like a corporation, then you can finally do one honest thing like a corporation:
No more subsidies.
Pay your taxes.
Brant Mittler retweeted
Suppose restaurants became tax exempt.
Five years later they held billions in assets.
Paid CEOs $33 million.
Reported hundreds of millions in annual surplus.
Then argued they deserved even larger taxpayer subsidies.
Would anyone still call them charities?
Why is healthcare different?
Brant Mittler retweeted
A physician submits a clean claim and gets denied.
A fraud ring submits fake DME claims and allegedly gets millions.
Incredible payment integrity system.
Brant Mittler retweeted
Physician conversion factors:
2000: $36.6173
2003: $36.7856
2006: $37.8975
2009: $36.0666
2012: $34.0376
2015: $35.7547
2018: $35.9996
2021: $34.8931
2024: $32.7442
UHC stock up 8,556%
Insurance premiums up 342%
Inflation up 86%
Physician pay down 10%
We should keep cutting physician pay so families have more money in their pockets through reduced costs and reduced premiums based on the last 25 years of experience.
Average US family health insurance premium...
2000: $6k
2003: $9k
2006: $11k
2009: $13k
2012: $16k
2015: $18k
2018: $20k
2021: $22k
2024: $26k
That's a 342% increase since 1999 (6.1% per year).
(Note: US CPI inflation has increased 2.5%/year)
youtube.com/channel/UCRoWRnX…
Brant Mittler retweeted
If you think all we need to do in healthcare is let people shop for prices and they will fall, is ridiculous
The number of hospitals and insurance comps walking away from each other, particularly for Med Adv, tell us everything
That the entire HC system is designed to make it IMPOSSIBLE to price shop
Hospitals don't know their costs and can't set prices to insurance companies
Insurance companies do their best to manipulate transactions (latelq, underpay, deny, etc)
All of MA is an attempt by the carriers at arbitrage They bet that they can break the law, and never adhere to reporting regs and the worse that can happen is they might get fined
They bet that hospitals don't know their margins or profits on a per plan or carrier basis and they will make stupid decisions That is starting to change
Now they are using AI to manipulate prices and costs in real time, knowing providers use consultants for Rev mgt, making them unable to respond in months, let alone real time
They bet that by gaming MLR with subsidiaries, they can further break the law and not get caught
Bottom line , it's not an efficient market , due to zero transparency at the transaction level ,and the concept that individuals can shop based on price , when neither hospitals or insurance companies know what their actual costs and prices are , is insane
Carriers manipulate and obfuscate every number they can with the goal of making it impossible to know actual prices
Good luck shopping for prices on all bu the simplest, most obvious services
Brant Mittler retweeted
Totally agree. And it’s also why so many of your local physicians are struggling right now. They’re struggling because they don’t want to work for these behemoths, and yet, they really can’t always compete. These systems generate millions and millions of dollars, and pour a lot of that into marketing.
This is why building a brand using the online space is so valuable. If you rack up followers, they’ll come see you regardless if you work for the local money machine health system.
It’s also why I created @somedocs - so that the public could find the individual physicians that run a health interaction, rather than just get swayed by who the system tells you to see. I built it to provide the publicity and marketing that our peers so badly need, especially those that break off from those systems to practice privately.
Doctorsonsocialmedia.com
Houston airport is covered in billboards for nonprofit health systems.
Handsome leaders.
White coats.
Academic prestige.
“Better health.”
Cute.
These are the same institutions that beg the legislature every year for taxpayer dollars while fighting competition, acquiring physicians, inflating prices, and pretending the word “nonprofit” means “public good.”
America does not need more healthcare slogans.
It needs fewer grifter institutions living off taxpayers while making care less affordable.
Brant Mittler retweeted
A study published today in Science may be the most important AI paper in clinical medicine this year. And it happened to land on the same day I submitted a letter to JAMA arguing that AI can already deliver clinically adequate care for defined tasks.
Researchers at Harvard Medical School and Beth Israel Deaconess Medical Center ran six experiments pitting OpenAI's o1 reasoning model against hundreds of physicians across the full spectrum of clinical reasoning: differential diagnosis, management planning, probabilistic reasoning, and clinical documentation. Then they did something most AI studies don't. They tested it on 76 real, unstructured emergency department cases pulled directly from the medical record at a major academic medical center.
The results across all six experiments: the AI outperformed physicians.
On the real ER cases — the messiest, most clinically relevant test — the AI identified the correct or very close diagnosis in 67.1% of cases at initial triage, 72.4% at ER physician evaluation, and 81.6% at hospital admission. The two attending physicians scored 55.3% and 50.0% at triage, 61.8% and 52.6% at ER evaluation, and 78.9% and 69.7% at admission. The gap was widest at initial triage.
On management reasoning using expert-scored clinical vignettes, the AI scored a median of 89%. Physicians with conventional resources scored 34%. That is not a typo.
The physician evaluators were blinded and could not distinguish AI-generated differentials from human ones. One evaluator guessed correctly 15% of the time. The other guessed correctly 3% of the time.
I'm an emergency physician. I work in a rural Texas ED. These are my cases. These are my decision points. And I can tell you that the triage finding is the one that matters most. Triage is where the least information meets the highest stakes — where the wrong call means a patient sits in the waiting room while their sepsis progresses or their STEMI evolves. The AI was 12 to 17 percentage points better than experienced attendings at exactly that moment.
The authors are careful to note this is text-based reasoning only; the AI doesn't see the patient's distress, doesn't hear breath sounds, doesn't read the room. Those are real limitations today. But the cognitive reasoning component of emergency medicine — pattern recognition under uncertainty with incomplete data — is precisely what this model is demonstrating it can do.
This was published in Science. Not a preprint. Not a company blog post. Peer-reviewed, in one of the two most prestigious scientific journals in the world.
The profession needs to stop debating whether AI will be good enough. It needs to start planning for the fact that, for an expanding set of clinical reasoning tasks, it already is.
And yes, this was written with AI. Sorry!!
Brant Mittler retweeted
Houston airport is covered in billboards for nonprofit health systems.
Handsome leaders.
White coats.
Academic prestige.
“Better health.”
Cute.
These are the same institutions that beg the legislature every year for taxpayer dollars while fighting competition, acquiring physicians, inflating prices, and pretending the word “nonprofit” means “public good.”
America does not need more healthcare slogans.
It needs fewer grifter institutions living off taxpayers while making care less affordable.
Brant Mittler retweeted
Medicare affects many aspects of physician care- especially payment.
I just sent this formal letter to the new HHS/CMS Healthcare Advisory Committee ([email protected]), Secretary Robert F. Kennedy Jr., and Administrator Oz.
After 30+ years as an independent neurologist, I had to speak up. Click on link below to see letter:
docs.google.com/document/d/1…
Brant Mittler retweeted
This article should be mandatory reading for every medical student, PhD candidate, researcher—and honestly, for anyone who mistakes expertise for certainty.
“The importance of stupidity in scientific research” sounds provocative, almost offensive. But Martin Schwartz is not glorifying incompetence. He is describing the real operating system of discovery.
Science is not built on knowing.
Science is built on tolerating not knowing.
That distinction matters.
Most of education rewards correctness.
School teaches us to answer.
Exams reward speed, certainty, and precision.
You feel intelligent when you get things right.
Research is the opposite.
Real research begins exactly where competence ends—at the frontier where nobody knows the answer, including the people you thought must know.
That moment is psychologically brutal.
You ask the expert.
The expert shrugs.
You assume you’re missing something.
Then you realize: no—this is the work.
You are not failing.
You are standing at the actual boundary of knowledge.
That feeling—“I must be stupid”—is often not a sign of inadequacy.
It is often the first sign that you are finally asking an important question.
Medicine struggles with this.
We train doctors to avoid uncertainty, to fear being wrong, to perform confidence.
But the best clinicians and the best scientists know how to sit inside ambiguity without collapsing into fake certainty.
This is why AI in medicine also deserves caution.
Systems trained only to reproduce established answers may become extraordinarily good at passing exams while being terrible at discovering what matters next.
Guideline intelligence is not the same as scientific intelligence.
Discovery requires productive stupidity:
the willingness to stay with the uncomfortable,
to look ignorant,
to ask naïve questions,
to be wrong repeatedly without protecting your ego.
Most people want the authority of expertise.
Very few want the humiliation required to earn it.
But progress lives there.
Not in certainty.
Not in performance.
Not in sounding smart.
In the quiet discipline of saying:
“I don’t know… yet.”
And continuing anyway.
Brant Mittler retweeted
In 2016, Geoff Hinton told a room in Toronto to stop training radiologists because AI would replace them within five years.
If anyone had listened, we would have produced zero new radiologists between 2016 and 2021.
UCSF cannot hire them fast enough today.
Robert Wachter, MD, chair of medicine at UCSF and author of A Giant Leap: How AI Is Transforming Health Care, has been thinking about what that misread actually means for the rest of us.
The lesson is not that Hinton was wrong about the capability. He was wrong about the timeline, and he was wrong about what radiologists would do with the new tools.
Here is the framework worth bookmarking.
Every prior healthcare technology, including the EHR, was foundational but not transformative. Digitizing notes did not change what being a doctor felt like. It just moved the handwriting into a screen and added thirty-two checkboxes.
The EHR also did something subtler. It turned medicine into a surveillance surface. The moment the note became electronic, every stakeholder who cared what you did as a physician could look over your shoulder in real time and build enforcing functions to make you do things. That is the real source of clinician burnout, not the software.
AI is going to be that on steroids.
The new tools will not just suggest diagnoses. They will suggest treatments. The values embedded in those recommendations, including whether to offer a $50,000 Alzheimer's drug that slows progression by three months with a 1 percent risk of brain bleed, will not be chosen by you. They will be chosen by your institution or your payer.
And the disruption we should be watching is not radiologist replacement. Wachter notes the same pattern is about to reach procedural specialties. Why does a $450,000 gastroenterologist perform the colonoscopy when the colonoscope itself is AI-enabled and the skill could plausibly be taught to someone who is not in the fifteenth year of post-college training? That is not a hypothetical. That is the next arbitrage.
The real story of AI in medicine is the slow institutionalization of decision support into a system where the physician no longer owns the recommendation, but still owns the liability.
And here is the part health systems should be planning around now. Wachter argues the biggest risk to clinical AI adoption is a single high-profile AI error that kills a patient and resets the field by five years. His reference point is Cruise, the GM-owned driverless car company that struck a woman in San Francisco, did not kill her, and was out of business within a year. Waymo survived because it spent years building a trust reservoir, cruise control, then automatic braking, then millions of supervised miles, before the steering wheel came off. Medicine has not built that reservoir yet. The tools now in physician hands, scribes and chart summarization, are the cruise-control phase. The diagnosis and treatment layer is the steering wheel.
Listen to the full conversation on The Podcast by KevinMD. Link in the replies.
Who do you think will set the values inside clinical AI first, health systems or insurers?
#ThePodcastbyKevinMD
Brant Mittler retweeted
📣 We are excited and thrilled to announce APOLLO, a healthcare system-scale multimodal temporal foundation model for virtual patient representations.
Trained on 25 billion clinical events from 7.2 million patients across 33 years and 28 modalities, APOLLO learns a unified atlas of medicine. Turning labs, notes, pathology images, medications, and diagnoses into coherent, computable longitudinal trajectories. APOLLO is disease-agnostic by design, a single model that learns the shared structure underlying human health and disease across every specialty, modality, and stage of care.
The possibilities are enormous: earlier risk prediction, treatment response modeling, clinical trial matching, biomarker discovery, and a new generation of agentic systems built on rich patient representations.
Read the pre-print: arxiv.org/pdf/2604.18570
Read our blog post about the work: linkedin.com/pulse/apollo-mu…
👏 🎉Huge congratulations to Andrew Zhang , @TongDing99, Sophia J. Wagner, and the rest of the team.
Brant Mittler retweeted
Here’s the uncomfortable question nobody is asking about AI in medicine 👇
We keep celebrating that AI can “match specialists” because:
✔️ it passes board exams
✔️ it follows guidelines
✔️ it produces specialist-like recommendations
And from that, a conclusion is emerging:
👉 If AI scales specialist knowledge, we may not need specialists anymore.
Sounds elegant. It’s also deeply flawed.
Because this logic hides a critical assumption:
that knowledge = reasoning = clinical care
It doesn’t.
Even the same paper admits that AI is still weak in:
✔️ context-dependent reasoning
✔️ longitudinal decision-making
✔️ integrating fragmented patient realities
That’s not a side note. That’s the whole game.
Medicine is not a board exam.
Real medicine is:
✔️ incomplete data
✔️ conflicting signals
✔️ evolving disease over time
✔️ decisions under uncertainty
r✔️ esponsibility for outcomes
👉 In other words: not a retrieval problem
So what are we doing?
We are benchmarking AI where it is strongest:
✔️ structured knowledge
✔️ standardized logic
✔️ guideline concordance
And then projecting those results onto:
✔️ messy reality
✔️ real patients
✔️ longitudinal care
That leap is… optimistic.
But let’s challenge the counter-argument too.
Yes — historically, specialization was driven by “cognitive limits.”
Yes — AI removes part of that constraint.
But was knowledge really the main bottleneck?
Or was it:
✔️ integrating context
✔️ managing uncertainty
✔️ making accountable decisions over time
Because those are not solved by AI.
And here’s the paradox nobody is discussing:
👉 The better AI gets at detecting and standardizing,
👉 the more edge cases, ambiguity, and borderline scenarios we will uncover
Which doesn’t eliminate specialists.
It increases the need for real expertise.
Especially in fields like imaging, where:
✔️ seeing more ≠ understanding more
✔️ detecting more ≠ deciding better
We are framing the future as: Generalist + AI vs Specialist
But the real divide is:
👉 Standardizable medicine vs Non-standardizable medicine
AI will dominate the first.
Humans will still carry the second.
So maybe the real question is not:
“Will AI replace specialists?”
But:
👉 “Are we overestimating AI because we are testing it on problems that are structurally aligned with its strengths?”
Because passing exams is easy.
Owning decisions, over time, in real patients?
That’s still medicine.
Brant Mittler retweeted
You can't hold hospital privileges without taking call.
Call isn't compensated.
That's not a calling. That's coercion.
Internal medicine physician Corinne Sundar Rao names something physicians have absorbed for decades without ever being able to articulate it cleanly: call is labor. Not a professional obligation. Not a rite of passage. Labor. And it has been either unpaid or minimally compensated for as long as it has existed.
The mechanism is worth understanding. To practice what you trained to do, you need hospital privileges. To keep hospital privileges, you take call. There is no opt-out. The work is tethered, undefined, and endless. One day runs into the next. You carry a full patient load after an overnight on call and you are expected to keep going.
No other profession works this way. Pilots have mandatory rest periods because fatigue at high stakes is a patient safety issue. The argument applies equally to surgeons making complex decisions after 30 hours without sleep.
The hospitalist model is the clearest proof that the old system stopped working. Nobody decided philosophically that hospitalists were a good idea. The system just broke down until it had to adapt. Defined shifts replaced open-ended obligation. The word hospitalist was first coined in a 1996 New England Journal of Medicine article, and the model has been expanding ever since.
The laborist model in obstetrics followed the same logic. High-volume hospitals recognized that asking one physician to manage clinic, elective surgeries, GYN, and overnight deliveries was not a sustainable structure. So they separated the labor floor from the rest of the work.
Call hasn't made that transition yet across most of medicine. And the cost is becoming visible. Physicians are quietly leaving. Not dramatically. They are going part-time, shifting to direct care or concierge models, or stepping away from clinical medicine entirely. Physician compensation accounts for only 8 to 8.6% of total U.S. healthcare costs. This is not a compensation problem. It is a structural one.
Call is not an infinite resource. It is labor. Labor must be compensated, defined, and protected with rest.
Corinne Sundar Rao on The Podcast by KevinMD. Listen link in the first comment.
🤖 Made with AI
Brant Mittler retweeted
I want to reflect on the broken culture of innovation in U.S. healthcare delivery—and its nefarious impact on the system.
We declare success prematurely.
We obscure failure.
And we move on without inquiry.
Within the delivery system, the same organizations whose basic and clinical researchers perform rigorous, thoughtfully designed controlled trials of medicines will tout the success of delivery and financing innovation without requisite evaluation.
In the broader innovation ecosystem, we too casually equate funding, valuations, and transactions with success—and we too often elevate and declare success without applying appropriate skepticism.
Many believe this culture of salesmanship and hucksterism is harmless.
But on the other side of it are patients whose care is not getting better and an industry with misplaced conviction about progress when there is none.
I’ve seen this firsthand throughout my career.
To make this less abstract, I’ll focus on my own area of work over the last decade: leading and operating Medicare Advantage plans and care delivery organizations.
It is a sector that has birthed many venture capital darlings.
But with the benefit of hindsight, many of those companies had clinical outcomes that were, at best, thin—
And whose primary innovation was exploiting arbitrage opportunities in risk adjustment.
And yet, in retrospect, these “innovators” were platformed.
They were celebrated on the stages of major national forums.
And while Medicare Advantage is now a space where some of these behaviors are more widely understood, it is not unique to Medicare Advantage.
Rather, some of the distortions we see there are a window into a broader cultural problem.
One that, in a different form, was on display in the case of Elizabeth Holmes and Theranos.
We might like to believe that was an exceptional story in exceptional circumstances.
I don’t.
I think we have many Theranoses among us—just less visible, and more socially accepted.
Why?
I believe there are two root cause.
First, expert and funding bias.
We believe that startups revered and validated by experts are worthy of celebration.
And if they are richly funded, we assume they have been appropriately vetted.
Commercial success—often driven by relationships and the favor economy—is prematurely equated to clinical success.
Second, optimism bias.
We have such a strong desire to believe that things are getting better that the first hint of progress is met with celebration—rather than the skepticism we typically reserve for results that seem too good to be true.
This is how many of us have convinced ourselves that American healthcare delivery is improving—while most consumers, patients, and clinicians tell us otherwise every single day.
So how do we move forward?
Brant Mittler retweeted
You spent years in training so you could justify a prescription to someone who has never seen a patient.
Bettina Reed has practiced family medicine for 33 years. She watched the system go from a 45-dollar visit with no middleman to prior auths on four-dollar medications and insurance reps overriding clinical decisions.
When a pharmacy director told her she could not send in a patient's Zoloft over a technicality, she sent it anyway. "My patient care is gonna continue."
Meanwhile, health insurance executives pull 20 to 30 million in annual salary while lobbying Congress with hundreds of millions. Providers absorb the fallout. Patients absorb the cost. And violence against health care workers keeps climbing.
The system will not fix itself. Reed says clinicians need to stop gliding along and start pushing back.
Episode is in the comments.
#HealthcareReform #ClinicianBurnout #PatientAdvocacy #KevinMD
🤖 Made with AI
Brant Mittler retweeted
Systematic literature reviews take 12-18 months to complete. Looks like AI is going to fully automate systematic reviews sooner than later.
SciSpace ( @scispace) just launched an autonomous AI agent that conducts a systematic literature review with a single prompt.
Go to scispace[.]com and run the following prompt:
"Conduct a systematic literature review on [your topic]"
SciSpace agent will generate research questions based on the PICO framework.
You can review these questions and edit them according to your specific requirements.
The agent will also draft screening criteria that you can edit according to your needs.
Then the agent asks you to select the databases you want to use and the date range for paper.
After this step, everything is fully automated.
The agent will search for papers in the relevant databases, it will combine and rerank the papers.
Then it will start the title and abstract screening and include the papers that meet the include criteria.
In the next step, it will download the full text of included papers and screen them followed by data extraction.
Based on the extracted data, it generates a complete systematic literature review and also a PRISMA diagram.
It will also give you a table of papers included along with the rational for including them.
The only thing that is keeping AI agents to fully automate systematic literature reviews fields is the papers behind paywalls.
Check out the agent at scispace[.]com and see if you find its review useful.
🤝 Paid partnership
Brant Mittler retweeted
We have a tendency to assume that when a system underperforms, the solution is to make it more efficient.
The IDR system under the No Surprises Act doesn’t underperform. It performs exactly as structured.
Insurers collect premiums upfront.
Physicians deliver care first and collect second.
Every day of payment delay is a day that money sits on an insurer’s balance sheet earning a return.
Multiply 150 days of average delay across 1.46 million annual disputes.
The question isn’t why the system is slow.
The question is why we assume slowness is the problem rather than the product…
buff.ly/8iNbqlI
Brant Mittler retweeted
🦔A researcher invented a fake eye condition called bixonimania, uploaded two obviously fraudulent papers about it to an academic server, and watched major AI systems present it as real medicine within weeks.
The fake papers thanked Starfleet Academy, cited funding from the Professor Sideshow Bob Foundation and the University of Fellowship of the Ring, and stated mid-paper that the entire thing was made up. Google's Gemini told users it was caused by blue light. Perplexity cited its prevalence at one in 90,000 people.
ChatGPT advised users whether their symptoms matched. The fake research was then cited in a peer-reviewed journal that only retracted it after Nature contacted the publisher.
My Take
The researcher made the papers as obviously fake as possible on purpose. The AI systems didn't catch it. Neither did the human researchers who cited it in real journals, which means people are feeding AI-generated references into their work without reading what they're actually citing.
I've covered the FDA using AI for drug review, the NYC hospital CEO ready to replace radiologists, and ChatGPT Health launching this year. All of that is happening in the same environment where a condition funded by a Simpsons character and endorsed by the crew of the Enterprise was being presented as emerging medical consensus. The people making these deployment decisions seem to believe the pipeline from research to AI to patient is more supervised than it actually is. This experiment suggests it isn't supervised much at all.
Hedgie🤗
nature.com/articles/d41586-0…
